Relationship between Molten Oxide Structure and ...

ISIJ International, Vol. 54 (2014), No. 9, pp. 2077–2083

Relationship between Molten Oxide Structure and Thermal Conductivity in the CaO–SiO2–B2O3 System Youngjae KIM* and Kazuki MORITA Department of Materials Engineering, Graduate School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656 Japan. (Received on January 30, 2014; accepted on April 30, 2014)

Using a transient hot-wire method, the thermal conductivity of the CaO–SiO2–B2O3 mold flux system was measured. The effects of temperature, BO1.5 concentration and basicity on the thermal conductivity were considered, along with structural investigation by Raman spectroscopy. It was found that the addition of boron oxide caused both a decrement and increment of thermal conductivity, depending on the basicity. These conflicting effects on thermal conductivity were considered to be caused by the following two different behaviors in the oxide melts. Boron oxide is incorporated into silicate networks at a lower basicity, while it tends to form borate networks at higher CaO/SiO2 ratios. In the case of basicity dependency, thermal conductivity initially decreases or remains constant with increasing CaO/SiO2 ratio in regions of low basicity, but increases when the CaO/SiO2 ratio is higher than 1.15. Due to the incorporated state of boron oxide in silicate networks at low basicity, the thermal conductivity is likely to be predominantly affected by the silicate networks. However, at a relatively high CaO/SiO2 ratio, an increase in chain-type metaborate was observed through Raman spectroscopy; this structural change in borate being responsible for the increment in thermal conductivity with higher basicity. Finally, the apparent activation energy of thermal conductivity was calculated, and was found to be reduced by the addition of boron oxide. KEY WORDS: thermal conductivity; transient hot-wire method; Raman spectroscopy; slag structure; mold flux.

sumed that this results from the weaker bond energy of B– O compared to Si–O or alkaline earth metal-oxygen bonds.11) In the blast furnace slag system, according to Ren et al.,20) even though the addition of boron oxide increases the amount of 4-coordinated boron oxide ([BO4]5–), the viscosity is nonetheless reduced by a lowering of the eutectic temperature. Similarly the complicated effects of B2O3 on its physical properties were also pointed out by Fox et al.18) They reported that because of the formation of three-fold and/or four-fold co-ordination of borosilicates, the effect of B2O3 was to create both a decrease and increase in viscosity for various mold flux systems. These previous works have mainly focused on viscosity, which is related to its lubrication of the mold flux. However, many common defects that occur during continuous casting processes, such as longitudinal cracking or sticker breakouts, result from either an excess or shortage of horizontal heat transfer.2,18) An understanding of the thermal conductivity of mold flux systems is therefore in high demand for efficient design and optimization of mold fluxes.9) It is, however, difficult to accurately measure the thermal conductivity of a molten slag, due to the large effect of convection and radiation at high temperatures.3) In order to reduce these effects, a transient hot-wire method has been adopted for measuring the thermal conductivity of molten slags at high temperature.9,10,21–27) In the silicate-based iron and

1. Introduction In the continuous casting process, the proper design and optimization of the mold flux is a significant concern in reducing the likelihood of various defects; as well as improving the surface quality of the final product.1,2) To date, it has been widely accepted that the structure of silicate systems, such as their degree of polymerization and the effect of cations, is directly related to the physical properties of molten slag.3,4) In this sense, studies pertaining to the relationship between the structure and physical properties of molten oxides have been carried out by a number of researchers with the intent to better design and optimize mold flux systems.5–13) Because of the poor application of the additive rule,14) and anomalous behavior in alkali/alkali earth oxide bearing systems,15,16) the effect of boron oxide in mold flux systems has been given particular attention.11–13,17–19) Wang et al.19) reported that the addition of B2O3 leads to an increase number of [BO4] tetrahedral units, which play an important role as network forming oxides in CaO–SiO2– B2O3–TiO2 slag systems. However, they also recently found that a decrease in viscosity occurs at higher B2O3 concentrations in CaO-SiO2-B2O3-TiO2-based systems, and pre* Corresponding author: E-mail: [email protected] DOI: http://dx.doi.org/10.2355/isijinternational.54.2077

2077

© 2014 ISIJ

ISIJ International, Vol. 54 (2014), No. 9

2.2. Thermal Conductivity Measurement The transient hot-wire method was adopted in this study for measuring the thermal conductivity of the molten slags. Figure 1 shows a schematic diagram of the experimental apparatus. Using a PID (proportional integral differential) controller and calibrated B-type thermocouple, sample temperature was controlled. The upper part of sample is intentionally placed at the highest temperature zone, in order to avoid natural convection during measurement. Approximately 90 g of pre-melted sample is heated in an electric resistance furnace equipped with SiC at 1 773 K, and then held for 1 hour to obtain a homogeneous molten slag. Thermal conductivity was measured at 50 K intervals from 1 773 K to the liquidus temperature. Using a galvanostat, a constant current of 0.8–1.5 A was supplied to a 0.15 mmφ Pt-13%Rh hot-wire, and any voltage change between the two terminals of the wire was monitored by a digital multimeter. A linear relationship between ΔV and lnt was obtained within 0.8–2.0 seconds, and the thermal conductivity, λ (W/m/K), was calculated by the following equation:

steelmaking slag system, it has often been reported that a more polymerized slag system shows a higher thermal conductivity, thus indicating a structural dependency.3,10,24–26) Kang and Morita23) measured the thermal conductivity in the CaO–SiO2–Al2O3 conventional iron-making slag system using the hot-wire method, reporting that the movement of phonons through covalent bonds is more effective than through ionic bonds in silicate melts; thus resulting in a higher thermal conductivity in more polymerized slag systems. In this study, the thermal conductivity of the CaO–SiO2– B2O3 mold flux system was measured in its molten state using the transient hot-wire method. As-quenched slag samples were also investigated by Raman spectroscopy, in order to elucidate any relationship between the structure of the molten oxide and its thermal conductivity. The apparent activation energy was also considered, which is derived from the slope of inverse temperature and evaluated from a structural point of view. 2. Experimental

λ=

2.1. Experimental Procedures Samples were prepared using reagent grade SiO2 and B2O3, and CaO calcined from CaCO3. The powder mixtures were pre-melted in a Pt crucible at 1 773 K for 30 minutes. After quenching, a finely crushed sample was used to fill a Pt-10%Rh crucible (I.D: 32 mm, O.D: 38 mm, height: 70 mm), and the thermal conductivity was then measured by the hot-wire method. Following measurement, the solidified samples were re-melted at 1 773 K, and then quenched. The resulting vitreous slag samples were ground and sieved to smaller than 100 μ mφ. The results of X-ray diffraction (XRD; RINT 2500, Rigaku, Japan) showed no characteristic peaks, thus confirming a non-crystalline state. The B2O3 and CaO contents of the final composition were analyzed by inductively coupled plasma atomic emission spectroscopy (ICP-AES; SPS3250UV, SII NanoTechnology, Japan), and the SiO2 contents were determined by a gravimetric analysis technique. The analyzed chemical composition is listed in Table 1. Finally, the structure of the melt was investigated by Raman spectroscopy.

Q 4π

d ΔV ............................. (1) dlnt

where Q(W/m), ΔV(V) and t(s) represent the heat generation rate of the hot-wire per unit length, voltage change and time, respectively. The derivation of Eq. (1) and further theoretical details have been described elsewhere.23,28) 2.3. Structural Investigation Structural investigation was carried out using Raman spectroscopy (T-64000, Horiba Jobin-Yvon, France), with excitation provided by an argon ion laser with the wavelength of 514 nm operated at 100 mW. The resulting Raman signal was collected in the wavenumber range of 400 to 1 600 cm–1. Although Raman spectroscopy provides information on molecular vibrations, the structural information cannot be obtained directly from the Raman spectra. Thus, in order to obtain this structural information, the spectra separating procedure was preceded by Gaussian deconvolution using the PeakFit V4 program. 3. Results and Discussion 3.1.

Table 1.

Effect of Temperature and Composition on Thermal Conductivity The thermal conductivities obtained from the CaO–SiO2– B2O3 slag system at various temperatures are shown in Fig. 2. From this, it can be seen that the thermal conductivity is gradually increased with lower temperature. In most ceramics systems, heat is mainly transmitted through quantized lattice vibration, known as phonon vibration.29) According to Ziman,30) thermal conductivity by phonons (λ ) can be expressed by following equation, which is analogous to kinetic theory:

Initial and final chemical compositions of present work.

CaO/SiO2 (molar ratio)

Initial composition (mol%)

Final composition (mol%)

CaO

SiO2

BO1.5

CaO

SiO2

BO1.5

1

0.92

45.5

49.5

5.0

46.2

48.4

5.4

2

0.92

40.7

44.3

15.0

41.0

43.0

16.1

3

0.92

35.9

39.1

25.0

34.7

38.1

27.3

4

1.15

50.8

44.2

5.0

51.5

43.1

5.4

5

1.15

45.5

39.5

15.0

45.0

37.8

17.2

6

1.15

40.1

34.9

25.0

39.0

35.9

25.1

7

1.38

55.1

39.9

5.0

55.6

39.0

5.4

8

1.38

49.3

35.7

15.0

48.1

36.3

15.6

9

1.38

43.5

31.5

25.0

42.7

32.1

25.2

© 2014 ISIJ

1 λ = Cvl ................................. (2) 3 where C is the total specific heat, v is the mean particle velocity, and l is the phonon mean free path of collision. It is reported that the total heat capacity (C) and mean particle velocity (v) become approximately constant for most ceram2078


Fig. 3. Fig. 1.

Experimental apparatus for thermal conductivity measurement.

Fig. 2.

Temperature dependence of thermal conductivity in the CaO–SiO2–B2O3 oxide system.

Relationship between thermal conductivity and BO1.5 concentration at a fixed CaO/SiO2 ratio.

Ammar et al.,32) an increase in the disorder of a network structure in silicate glass results in a shortening of the phonon free path. Mills3) also suggested that the movement of phonons along the silicate chain, or ring, is relatively easier than from chain to chain. Therefore, the addition of boron to a CaO–SiO2 based slag system is believed to increase the mean free path of the phonons by changing the network structure of molten oxides. The effect of boron oxide concentration on thermal conductivity is shown in Fig. 3. From this, it can be seen that the addition of boron oxide has opposite effects on thermal conductivity depending on the CaO/SiO2 ratio. When this ratio is 0.92, the thermal conductivity is reduced at higher boron oxide concentrations. However, at higher ratios of 1.15 and 1.38, the thermal conductivity either remains unchanged or increases slightly with increasing boron oxide concentration. Fox et al.18) reported that such a conflicting effect of boron oxide on the physical properties is due to the nature of boron coordination, and that similar behavior should be observed in the case of the viscosity. Figure 4 shows the change in thermal conductivity as a function of CaO/SiO2 ratio at a fixed concentration (BO1.5). In all cases, the thermal conductivity decreases or remains constant with an increase in basicity when the CaO/SiO2 ratio is smaller than 1.15, whereas it increases with increasing basicity when the CaO/SiO2 ratio is greater than 1.15. Teixeira et al.33) reported that the activity coefficient of BO1.5 in a CaO–SiO2 binary slag system increases with basicity in acidic slag compositions, while decreasing under basic conditions. They explained that at a molar ratio of CaO/SiO2 less than 0.86, boron oxide is incorporated into the silicate network. On the other hand, B2O3 behaves as an acidic oxide under conditions of higher basicity, thus resulting in a decrease in the activity coefficient of boron oxide with increasing CaO/SiO2 ratio. Although Teixeira’s works were carried out at relatively low boron oxide concentrations, their results are nonetheless worthy of qualitative consideration. From a thermodynamic point of view, it can be

ics near and above the Debye temperature (100–1 000 K).29) At higher temperatures, the thermal conductivity would therefore be mainly affected by its mean free path (l), which is proportional to the inverse of temperature (1/T).29,30) It can therefore be considered that the observed temperature dependency of the thermal conductivity results from a decrement of the mean free path (l) with higher temperature, as described by Eq. (2). Compared to the thermal conductivity of a CaO–SiO2 binary system, as reported by Nagata and Goto,31) the 1.38 CaO/SiO2 ratio of the present study produces a relatively higher thermal conductivity above 1 700 K, which is approximately equivalent to the liquidus temperature of the CaO–SiO2 slag system. According to 2079

© 2014 ISIJ


CaO/SiO2 ratio of 0.92, the addition of basic oxide leads to a depolymerization of the silicate structure and a resulting decrease in thermal conductivity. On the other hand, under more basic conditions, it is likely that the change in borate structure also effects the change in thermal conductivity. More detailed structural investigation will be discussed in the following sections.

inferred that owing to the acidic behavior of boron oxide, the addition of boron oxide should lead to polymerization of the silicate resulting in an increase in thermal conductivity with higher BO1.5 concentrations at CaO/SiO2 ratios of 1.15, 1.38. With a ratio of 0.92 the boron oxide is incorporated into the silicate structure, thus B2O3 addition would change the nature of this network structure and disrupt the movement of phonon, thereby reducing its thermal conductivity. Such variation in the boron oxide state with slag basicity would therefore explain the vastly different effects of B2O3 addition on thermal conductivity. Similarly, the basicity dependency of thermal conductivity is related to the change of state of boron oxide. Since boron oxide is incorporated into silicate networks at low basicity, thermal conductivity is mainly affected by the resulting change of silicate structure. Consequently, at a

Fig. 4.

3.2.

Investigation of Structure by Raman Spectroscopy and its Effect on Thermal Conductivity Figure 5 shows the Raman spectra and related Gaussian deconvoluted bands of non-crystalline CaO–SiO2–B2O3 within a 400 to 1 600 cm–1 range of Raman shift. These Raman spectra have been identified from appropriate references34–41) and are listed in Table 2. Owing to the complicated structure of boron oxide, the possibility of a borate structure existing within the CaO–SiO2 system was investigated; and an annular metaborate group, chain-type metaborate group, pyroborate group and six-membered borate ring were all considered in the present study. It should be noted that a danburite structure of CaO•B2O3•2SiO2 (one BO4– unit is surrounded by three SiO4 and one BO4–) was not considered which is observed around 630 cm–1 of Raman shift.40) According to Osipov et al.,42) a danburite spectrum is observed at 614 cm–1 with a complete absence of peaks within the 1 400–1 500 cm–1 range, however, spectra within the range were found almost the entire compositions in the current study. In addition, a fully-polymerized tetrahedral silicate (SiO4) unit, which is an essential component for danburite, was rarely observed. Consequently, the Raman spectra around 600–650 cm–1 were assigned to ring-breathing vibrations of ring-type metaborate units instead of danburite. The integration of Gaussian deconvoluted spectra of silicate structural units, and their relative fractions, are summarized in Fig. 6. Qn species indicate the SiO4 tetrahedron with n of bridging oxygen and (4-n) of non-bridging oxygen. It is worth noting that the silicate structure is more polymerized with a higher boron oxide concentration, or a lower CaO/SiO2 ratio. This polymerization tendency was observed

Relationship between thermal conductivity and CaO/SiO2 ratio at a fixed BO1.5 concentration.

Fig. 5.

© 2014 ISIJ

Raman spectra and its Gaussian decovolution bands of the non-crystalline CaO–SiO2–B2O3 samples with 5, 15, 25 mol% of BO1.5 at a fixed CaO/SiO2 ratio of (a) 0.92, (b) 1.15, (c) 1.38.

2080

ISIJ International, Vol. 54 (2014), No. 9 Table 2.

Reference peak positions of Raman and its structures in the CaO–SiO2–B2O3 system.

Reference position (cm–1)

Assignments

490–570

“Isolated” diborate units34,35)

600–650

Ring-breathing vibration of ring-type metaborate34–37)

700–730

B–O–B stretching in chain metaborates36,37)

750–780

Breathing vibration of a six membered ring containing both BO3 triangles and BO4 tetrahedra34,35,38)

850

Symmetric stretching vibrations of tetrahedral silicate units with 4 non-bridging oxygens39)

900


950–1 000


1 050–1 100


1 150

Asymmetric silicon-oxygen stretching vibrations within a fully-polymerized tetrahedral silicate network39)

1 200–1 300

Symmetric stretching vibration of terminal B–O– bonds in pyroborate unit36–38)

1 410

Stretching of B–O– bonds attached to BO4 unit40,41)

1 480

Stretching of B–O– bonds attached to BO3 unit40,41)

Fig. 6.

Relationship between the relative fraction of silicate structure and (a) mol% of BO1.5 at fixed CaO/SiO2 ratio of 1.15, and (b) CaO/SiO2 ratio at a fixed 15 mol% of BO1.5.

tration, using MAS-NMR analysis. It can therefore be considered that the increment of thermal conductivity with a higher BO1.5 concentration, at CaO/SiO2 ratios of 1.15 and 1.38, results from the polymerization of a network structure supporting the acidic behavior of boron oxide. However, in the case of a CaO/SiO2 ratio of 0.92, the thermal conductivity is reduced in spite of this polymerization due to the increase in boron oxide concentration. As explained in section 3.1, this may be related to the incorporated state of boron oxide in silicate network structures. It seems likely that an increase of BO1.5 concentration results in a larger polymerized borate structure incorporated in the silicate structure, thus contributing to a decrease in thermal conductivity by disrupting the movement of phonons. The relative borate structural fraction with varying CaO/ SiO2 ratio at a fixed BO1.5 concentration is shown in Fig. 8. This demonstrates the decrease in ring-type metaborate, and related increase in chain-type metaborate, at CaO/SiO2 ratios of more than 1.15. Since boron oxide is incorporated into the silicate network structure at low basicity, the depen-

in every composition of this study, and the results are therefore in accordance with a previous structural study by MAS-NMR spectroscopy.17) However, compared to these MAS-NMR results, the absolute value of the relative fraction of each silicate structure does show some discrepancy. In the MAS-NMR analysis of Angeli et al.,43) the effect of Ca ions on 29Si chemical shift is more significant than silicate polymerization; and for this reason, they considered Raman spectroscopy as a complementary structural data tool. The discrepancy in the relative fractions of the silicate structure obtained by Raman spectroscopy and MAS-NMR would therefore likely be caused by the effect of the second neighbor’s nature on 29Si chemical shift in MAS-NMR. In Fig. 7, the addition of BO1.5 leads to both increase in the stretching vibration of B–O– attached to BO4 and BO3 units and decrease in ring-type metaborate units, regardless of the basicity. It can be inferred that the structure of molten slag becomes polymerized with increasing boron oxide concentration. Sakamoto et al.17) also reported a similar increment of 4-coordinated boron with higher B2O3 concen2081

© 2014 ISIJ


Fig. 7.

Relationship between the relative fraction of borate structure and mol% of BO1.5 at fixed CaO/SiO2 ratio of (a) 0.92, (b) 1.15, and (c) 1.38.

Fig. 8.

Relationship between the relative fraction of borate structure and CaO/SiO2 ratio mol% at fixed mol% of BO1.5 (a) 5 mol%, (b) 15 mol%, and (c) 25 mol%.

of transport.44) In the case of viscosity, the temperature dependency is simply described by the Arrhenius equation; from which the apparent activation energy can be derived. As an analogy to viscosity, Kang and Morita23) expressed thermal conductivity as an Arrhenius-type equation, based on the relationship between the mean free path and the number of thermally broken bonds. Using this equation, values for the apparent activation energy of thermal conductivity were obtained, and are plotted in Fig. 9. The activation energy of the present slag system was between 91 kJ/mol to 243 kJ/mol, which is similar values compared to other B2O3 bearing slag systems.11,12) The results of the present study show that the activation energy of thermal conductivity decreases with a higher BO1.5 concentration, indicating an increment in the number of thermally broken bonds or a decrement in the required energy to break oxygen bonds. This result seems to be contradictory to the network formation with higher boron oxide concentration explained in section 3.2. However, Wang et al.11) measured the viscosity in the B2O3 and TiO2 bearing mold flux system, reporting the decreasing of activation energy with polymerization due to the increase in boron

dency of thermal conductivity on CaO/SiO2 ratio is mainly affected by this change in the silicate network structure. An initial decrease or lack of change in the thermal conductivity with higher CaO/SiO2 ratios therefore results from depolymerization of silicate structure, as-supported by the results of Raman spectroscopy. On the other hand, at CaO/SiO2 ratios of greater than 1.15, the thermal conductivity increases with higher basicity as a result of an increment of chaintype metaborate, which enhances the movement of phonons by lengthening the network structure. However, it is still unclear as to why this increment of chain-type metaborate has a greater influence over the thermal conductivity than the depolymerization of the silicate structure. More thorough investigation incorporating both thermodynamic and structural aspects is therefore needed to elucidate these conflicting effects of borate and silicate structure on thermal conductivity. 3.3.

Relations between Activation Energy and its Molten Structure In silicate melts, the energy required to break oxygen bonds can be determined by measuring the activation energy © 2014 ISIJ

2082


Acknowledgement The authors greatly appreciate the helpful advice on Raman spectroscopy measurement provided by Prof. T. Yoshikawa and Dr. S. Kawanishi at The University of Tokyo. REFERENCES

Fig. 9.

1) K. C. Mills, A. B. Fox, Z. Li and R. Thackray: Ironmaking Steelmaking, 32 (2005), 26. 2) K. C. Mills and A. B. Fox: ISIJ Int., 43 (2003), 1479. 3) K. C. Mills: ISIJ Int., 33 (1993), 148. 4) J. D. Frantza and B. O. Mysen: Chem. Geol., 121 (1995), 155. 5) W. McCauley and D. Apelian: Can. Metall. Q., 20 (1981), 247. 6) Z. Zhang, G. Wen, P. Tang and S. Sridhar: ISIJ Int., 48 (2008), 739. 7) M. Persson, M. Görnerup and S. Seetharaman: ISIJ Int., 47 (2007), 1533. 8) R. G. Hill, N. D. Costa and R. V. Law: J. Non-Cryst. Solids, 351 (2005), 69. 9) S. Ozawa and M. Susa: Ironmaking Steelmaking, 32 (2005), 487. 10) M. Susa, S. Kubota, M. Hayashi and K. Mills: Ironmaking Steelmaking, 28 (2001), 390. 11) Z. Wang, Q. Shu and K. Chou: Steel Res. Int., 84 (2013), 766. 12) G. H. Kim and I. Sohn: Metall. Mater. Trans. B, 45 (2014), 86. 13) G. Li, H. Wang, Q. Dai, Y. Zhao and J. Li: J. Iron Steel Res. Int., 14 (2007), 25. 14) T. Nanba, S. Sakida and Y. Miura: Proc. Materials Science & Technology 2006, Cincinnati, USA, (2006), 535. 15) Y. Yun and P. Bray: J. Non-Cryst. Solids, 30 (1978), 45. 16) W. J. Dell, P. J. Bray and S. Z. Xiao: J. Non-Cryst. Solids, 58 (1983), 1. 17) M. Sakamoto, Y. Yanaba, H. Yamamura and K. Morita: ISIJ Int., 53 (2013), 1143. 18) A. Fox, K. Mills, D. Lever, C. Bezerra, C. Valadares, I. Unamuno, J. Laraudogoitia and J. Gisby: ISIJ Int., 45 (2005), 1051. 19) Z. Wang, Q. Shu and K. Chou: ISIJ Int., 51 (2011), 1021. 20) S. Ren, J. Zhang, L. Wu, W. Liu, Y. Bai, X. Xing, B. Su and D. Kong: ISIJ Int., 52 (2012), 984. 21) S. Ozawa, R. Endo and M. Susa: Tetsu-to-Hagané, 93 (2007), 416. 22) K. Nagata, M. Susa and K. Goto: Tetsu-to-Hagané, 69 (1983), 1417. 23) Y. Kang and K. Morita: ISIJ Int., 46 (2006), 420. 24) Y. Kang, K. Nomura, K. Tokumitsu, H. Tobo and K. Morita: Metall. Mater. Trans. B, 43 (2012), 1420. 25) M. Hayashi, H. Ishii, M. Susa, H. Fukuyama and K. Nagata: Phys. Chem. Glasses, 42 (2001), 6. 26) M. Susa, M. Watanabe, S. Ozawa and R. Endo: Ironmaking Steelmaking, 34 (2007), 124. 27) B. Glaser and D. Sichen: Metall. Mater. Trans. B, 44 (2013), 1. 28) M. Susa, K. Nagata and K. Goto: Trans. Jpn. Inst.. Met., 29 (1988), 133. 29) W. D. Kingery: Introduction to Ceramics, John Wiley & Sons, New York, (1967), 486. 30) J. M. Ziman: Electrons and Phonons: the Theory of Transport Phenomena in Solids, Oxford University Press, London, (1960), 288. 31) M. Susa and K. C. Mill: Slag Atlas, Verlag Sthleisen GmbH, Düsseldorf, (1995), 596. 32) M. Ammar, S. Gharib, M. M. Halawa, K. El Badry, N. Ghoneim and H. El Batal: J. Non-Cryst. Solids, 53 (1982), 165. 33) L. A. V. Teixeira, Y. Tokuda, T. Yoko and K. Morita: ISIJ Int., 49 (2009), 777. 34) G. Padmaja and P. Kistaiah: J. Phys. Chem. A, 113 (2009), 2397. 35) E. Kamitsos, M. Karakassides and G. D. Chryssikos: J. Phys. Chem., 91 (1987), 1073. 36) R. K. Brow, D. R. Tallant and G. L. Turner: J. Am. Ceram. Soc., 79 (1996), 2410. 37) H. Li, Y. Su, L. Li and D. M. Strachan: J. Non-Cryst. Solids, 292 (2001), 167. 38) B. Dwivedi and B. Khanna: J. Phys. Chem. Solids, 56 (1995), 39. 39) P. McMillan: Am. Mineral., 69 (1984), 622. 40) D. Manara, A. Grandjean and D. Neuville: Am. Mineral., 94 (2009), 777. 41) R. Akagi, N. Ohtori and N. Umesaki: J. Non-Cryst. Solids, 293 (2001), 471. 42) A. Osipov, L. Osipova and V. Eremyashev: Glass Phys. Chem., 39 (2013), 105. 43) F. Angeli, T. Charpentier, D. De Ligny and C. Cailleteau: J. Am. Ceram. Soc., 93 (2010), 2693. 44) B. Mysen and P. Richet: Silicate Glasses and Melts; Properties and Structure, Elsevier Science, Amsterdam, (2005), 116.

Change of activation energy with varying mol% of BO1.5 at different CaO/SiO2 ratio.

oxide concentration. They explained that the formation of weaker B–O bonds in the network structure leads to a lowering of the activation energy. It is therefore concluded that the increase of B–O bonds in the network structure caused by the addition of BO1.5 to a CaO/SiO2 system results in a weakening of the network bond energy, and thus a decrease in the activation energy. 4. Conclusion Using the transient hot-wire method, the thermal conductivity of the CaO–SiO2–B2O3 system was measured across a temperature range of 1 773 K to its liquidus temperature. It was found that the thermal conductivity was reduced with higher temperatures for every compositional range tested, and is the result of a decrease in the mean free path (l). However, the addition of boron oxide was found to result in both an increment and decrement of thermal conductivity, depending on the CaO/SiO2 ratio. At a relatively low CaO/ SiO2 ratio of 0.92, B2O3 is incorporated into silicate networks and higher concentrations therefore lead to reduced thermal conductivity by disrupting the movement of phonons. However, at higher basicities of 1.15 and 1.38, the thermal conductivity is increased at higher BO1.5 concentrations through the polymerization of both borate and silicate networks. It was also found that the thermal conductivity was initially decreased or remained constant with an increase in basicity, but was increased with CaO/SiO2 ratios of greater than 1.15. It was thought that at a relatively low basicity, boron oxide incorporated in the silicate structure and depolymerization of silicate structure at higher CaO/SiO2 ratios would mostly have an effect on thermal conductivity. Contrarily, it was found that with a higher basicity the thermal conductivity is in fact increased by an increment of a chain-type metaborate structure. A decrement of activation energy with increasing BO1.5 concentration was also observed, indicating a weakening of the bond energy by increasing of B–O bonds in the network structure of BO4 and BO3 units.

2083

© 2014 ISIJ